the application research of national ......months each year. the 2014-2015 0.5m resolution aerial...

THE APPLICATION RESEARCH OF NATIONAL GEOGRAPHY CENSUS DATA IN

THE DEPARTMENTAL INVESTIGATION AND MANAGEMENT-TAKING LAND

MANAGEMENT AS AN EXAMPLE

Na Jiang

Shandong Provincial Institute of Land Surveying and Mapping, 250102, Jinan, Shandong, China

[email protected]

Commission III, WG III/7

KEY WORDS: National Geography Census Data, Land Management, Land Cover Classification Data, Spatial Transfer Matrix,

Geographic National Conditions Monitoring, Yellow River Delta

ABSTRACT:

According to the "Natural priority, Status quo priority" principle of acquisition, the national geography census data has the

characteristics of objectivity, impartiality and accuracy. It provides a new perspective for the management and decision-making

support of other industries as a "third party" and plays an important role in the professional management and investigation of various

departments including land, transportation, forestry and water conservancy. Taking land resources supervision as an example, the

Yellow River Delta efficient eco-economic zone as the research area, based on the national geographic census data and the land

survey data, this paper established the correspondence of the two types of data through the reclassification of the land cover

classification data, calculated the spatial coincidence rate of the same land class and the circulation relations among different land

classes through the spatial overlay analysis and the calculation of space transfer matrix, quantified the differences between the data

and objectively analysed the causes of the differences; On this basis, combined with land supervision hot spots, supplemented by

multi-source remote sensing images and socio-economic data, analysed the application of geographic census data in the land

regulation from multi-point.

1. INTRODUCTION

Geographical national conditions reflect the natural, economic,

and cultural information of a country from the spatial

perspective(Ning and Wang, 2014), involving the territory of

the country, geographical divisions, topographic features, road

traffic networks, distribution of rivers, lakes and seas, land use,

town layout, spatial layout of productivity, etc. It is a

comprehensive expression of the natural and human geographic

information at the macro level (Chen, 2014; Li et al, 2012). In

order to comprehensively investigate and accurately express the

status quo of geographic national conditions, the national

geographical census data is divided into three kinds: surface

morphology, land cover classification, and the important

geographic national conditions elements. According to the

principle of "natural priority, status quo priority", the data has

the characteristics of objectivity, impartiality and high precision.

National geography census data has important differences from

departmental thematic data which is concerned about the

content of normal operation and social attributes. It provides a

new perspective for the management and decision-making

support of other industries as a "third party" and plays an

important role in the professional management and investigation

of various departments including land, transportation, forestry

and water conservancy.

With the rapid economic and social development, land use is

constantly changing and land resource is becoming increasingly

tense. The land administration has continued to strengthen the

supervision of land, and the monitoring of land resource has

been a research hotspot. Bao et al(Bao, 2003; You, 2009; Zhang

et al, 2012; Yang et al, 2014) extracted land use information

from remote sensing images for dynamic monitoring through

different information extraction methods; Zhang et al (Zhang

and Xiong, 2008; Zhong et al, 2009; Cao et al, 2013)combined

remote sensing and GIS technologies to analyze the spatial

structure changes of land use; Wan et al (Wan et al, 2005; Qiao

et al, 2013) monitored land use evolution based on land use

transfer matrix. Some of the researches that have been

conducted used low resolution image as the data source in large

area monitoring, which were difficult to meet the requirements

of fine-scale land use supervision; or they only monitored based

on the interpretation of remote sensing, and paid less attention

to the social attributes of land. In the actual land resource

supervision, it is necessary to use professional data of the land

department as the management basis, and the high-precision

"third-party" data is also required as the supplement to

supervise the results of land surveys. Taking this as a starting

point, this paper was based on national geography census data

and land use survey data, through the comparison and analysis

of two types of data, to clarify the relationship of data, verify

the characteristics of national geography census data and

explore the application and service direction. On this basis, it

could promote the in-depth application of national geography

census data in the investigation and management of land and

other sectors.

2. RESEARCH METHOD

The method of research in this paper is described as follows.

2.1 Reclassification of land cover classification data

First, we analysed the relationship between geographical

national surveys and thematic change surveys. Through the

analysis and comparison of national geography census data and

The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-3, 2018 ISPRS TC III Mid-term Symposium “Developments, Technologies and Applications in Remote Sensing”, 7–10 May, Beijing, China

This contribution has been peer-reviewed. https://doi.org/10.5194/isprs-archives-XLII-3-651-2018 | © Authors 2018. CC BY 4.0 License.

651

land use survey data, including principles of the classification,

differences of the contents and the spatial coordinate systems,

this paper determined the correspondence between the two

types of data and reclassified the national geographic census

data to the land use data system. Here, the land cover

classification data (hereinafter referred to as LCA) in the

national geography census data and the land parcel data

(hereinafter referred to as DLTB) in land survey data were used

to analysis. According to the corresponding relationship, the

LCA was reclassified to the land use classification system.

2.2 Data consistency analysis

Coordinate the spatial datum of the two types of data, and

convert the land use data from 1980 Xi'an coordinate system to

the 2000 national geodetic coordinate system; Perform spatial

overlay and statistical analysis, and calculate the spatial

coincidence rate of the same type of land.

2.3 Spatial transfer matrix analysis

The land use change transfer mode (Zhang, 2008) was used to

calculate the source-direction matrix from LCA to DLTB， and

to analyse the reason of discrepancy. To ensure the speed of

calculation, the two types of data were rasterized before

analysed.

2.4 Application in land use and supervision

Through the in-depth analysis of the two types of data

relationships, combined with the hot spots of land

administration, supplemented by multi-source remote sensing

images and socio-economic data, etc., the application of

national geography census data in land supervision was

analysed from multiple aspects.

The technical flow chart is shown in Figure 1.

Multi-phase

remote sensing

image data

LCA DLTB

Correspondence relationship

of two types of dataReclassify

Coordinate

transformation

Rasterize

Calculate transfer matrix

Source -

Directions

Analysis

Difference

analysis

Land Supervision

Application

Social,

economic data,

etc.

Rasterize

Figure 1. Technical flow chart

3. EXPERIMENTS AND ANALYSIS

3.1 Study area and data

The study area in this paper is the Yellow River Delta High

Efficient Ecological Economic Zone (hereinafter referred to as

the Yellow River Delta). The Yellow River Delta is an

economic zone formed by expanding to the periphery based on

the historical alluvial plain of the Yellow River and the coastal

areas of northern Shandong Province. It includes all the

Dongying and Binzhou cities and their adjacent areas which

have the similar natural and environmental conditions, such as

Hanting district, Shouguang city, Changyi city in the northern

part of Weifang; Laoling, Qingyun County in Dezhou, Gaoqing

County in Zibo, and Laizhou city in Yantai. The research area

involves a total of 19 counties (cities, districts) in 6 cities, with

a total area of 26,500 square kilometers, accounting for one-

sixth of Shandong province. The Yellow River Delta is one of

China's three major deltas and an important part of the Bohai

Rim economic circle. In 2009, the Yellow River Delta became a

national strategy and one of the main battlefields for the rapid

development of efficient ecological economy. However, the

Yellow River Delta's eco-environment is relatively fragile, and

the land use change is fast. Therefore, it is of great importance

to effectively supervise the land use in the area, so this paper

takes the area as the research area.

The national geography census data used in the experiment was

produced in year 2015 and mainly included land cover

classification data and geographic national conditions elements

data. Among them, the land cover classification data (LCA) has

about 1.83 million map spots, including 10 first-class categories,

which are cultivated land (0100), garden land (0200), forest

land (0300), grassland (0400), house building district (0500),

roads (0600), structures (0700), artificial excavation sites

(0800), deserts and exposed surfaces (0900), and water areas

(1000). On this basis, they are subdivided into 46 secondary

classes and 79 tertiary classes. The data uses the 2000 national

geodetic coordinate system, latitude and longitude coordinates.

Land use data was extracted from the 2014 land change survey

data. It was based on the land use survey results and initial

cadastral results, and updated based on the annual changes in

land registration and land statistics to meet the demands of land

management. The land use data of the Yellow River Delta came

from the Ministry of Land and Resources, with a total number

of 986,000 land parcels. The data is divided into 8 categories

such as cultivated land (01, hereinafter referred to as CLand),

garden land (02, hereinafter referred to as GLand), forest land

(03, hereinafter referred to as FLand), grassland (04), urban

villages and industrial land (20, hereinafter referred to as

ULand), transportation land (10, hereinafter referred to as

TLand), water areas and water conservancy facilities land (11,

hereinafter referred to as WLand) and other land (12,

hereinafter referred to as OLand). They are further divided into

30 second-class categories according to their operating

characteristics, utilization patterns, and coverage characteristics.

The data used 1980 Xi'an coordinate system, Gauss-Kruger 3

degree projection, and took 120º as the central meridian. In

addition, the 2010-2015 Land change survey remote sensing

image data was used as the base map for the analysis. Most of

these images were multi-spectral data with 0.5-meter spatial

resolution, partially 0.5-meter panchromatic imagery or better

than 2 meters multi-spectral data. The acquisition time is 10-12

months each year. The 2014-2015 0.5m resolution aerial

photography images and 2013-2015 2m resolution multi-

spectral data of Mapping Satellite-1, Resources Satellite-3 were

used as supplementary analysis data.

3.2 Built up the correspondence of data systems

Since the LCA and DLTB are independent data, the two types

of data do not correspond one by one. Therefore, in the process



652

of data analysis, the corresponding relationship must be

established first. In the process of establishing correspondence

relationship, according to the principles of “consistent in

content meaning, consistent in spatial reference, and consistent

in data statistical methods”, with the reference of land use data,

the land cover classification data was reclassified into the land

use classification data system, as shown in Table 1 which took

arable land as an example. Full coverage method was used to

complete the comparison of the whole classification system. For

the classes that could not be matched one-by-one, according to

the "major correspondence" principle, encoded them as the

mainly related class. The area of all categories should not be

recalculated or omitted. The meanings of classes with the same

name in the two classification systems may not be exactly the

same, and the correspondence relationship should be based on

the actual meaning of the classes. At the same time, the

correspondence relationship was based on the consistency

between subclasses. Upper level were merged hierarchically

according to subclasses.

Land Use Data Classification

Corresponding National

Geography Census Data

Classification

First Class Secondary

class First Class Secondary class

Code Name Code Name Code Name Code Name

01 Cultivated

Land

011 Paddy

Field 0100

Cultivated

Land 0110

Paddy

Field

012/

013

Irrigated

land /

Dry

Land

0100 Cultivated

Land 0120 Dry Land

0700 Structure 0750 Greenhouse

Table 1. Correspondence of two types of data (cultivated land)

3.3 Experiment analysis

3.3.1 Data difference analysis: Calculate the area of each

primary class in the DLTB data and calculate the area of the

first class of land use data which was reclassified from LCA

data. The land use data was transformed to the 2000 national

geodetic coordinate system, and the spatial overlay analysis of

the two types of data was performed in ArcGIS. Calculate the

ratio of the same type of overlapping area to the first class of

land use data.

The corresponding relationship between the two types of data is

shown in Figure 2. The overall coincidence rate was 62%.

Among them, CLand had the highest and most stable

coincidence rate, about 78%; followed by ULand, about 72%.

OLand had the lowest coincidence rate, only 6%. From the

perspective of spatial location, the coincidence rate raised from

the coast to the Inland, as shown in Figure 3. There are four

main reasons for the existence of differences. The first reason is

the difference in data collection principles. The national

geography census focus on natural attributes, while land use

surveys focus on management and social attributes. For example,

due to seasonal changes, economic benefits of planting, short-

term use of land contractors, etc., land use may be temporarily

changed, but land use data classes do not change; but In the

national geography census, according to the status of the natural

surface coverage at the time of its collection, it was assigned to

the corresponding category, as shown in Figure 4. The second

reason is that there are differences in the definitions of the same

class in national geography census and land survey. Take the

road definition as an example. In the national geography census,

road pavements with the width of more than 3 meters and the

length of more than 500 meters should be collected. In the

absence of vegetation coverage, this contains the range of the

embankment and the cut. The transportation land in land survey

includes “embankments, cuts, road ditch, bridges, car stops,

forests, and ancillary land directly served for the road” in the

design. As shown in Figure 5, the road widths in the national

geography census are mostly less than the road widths in the

land use surveys. The third reason is that the data collection

scales and the comprehensive principles are inconsistent. The

strips such as grassland, road pavement, and impervious surface

that meet the acquisition criteria in the land cover data may be

incorporated into the nearby category in land use, as shown in

Figure 6. The fourth reason is the inconsistent of data collection

date. The acquisition time of land use data is 2014, and the land

cover data is 2015. The inconsistency of data time results in the

change of actual land features. In addition, there is no one-to-

one correspondence between the two types of data, and there are

various relationships such as one-to-many, many-to-one, and

many-to-many correspondences. All of these cause differences

in the statistical results of the two types of data.

Figure 2. Corresponding relationship between the two types of

data

Figure 3. The overall coincidence in the Yellow River Delta



653

(a)

(b)

Figure 4. Inconsistent collection of the same object in two types

of data （a）Construction land in DLTB, grassland in LCA; the

red lines are the boundaries; the green points are the sample

points (b) Field photos of sample points

Figure 5. Inconsistent definition of road range in land use and

land cover

Figure 6. The impervious surface (the orange slash part) in LCA

was integrated in the DLTB

3.3.2 Source - Direction Analysis: On the basis of the

qualitative analysis of the two types of data relationships, the

conversion relationship between the data is quantitatively

calculated by source-direction analysis. In order to improve the

calculation efficiency, the two types of data were rasterized with

a grid size of 2 meters, and the source-and-direction matrix was

calculated in the software of ArcGIS. Take Dongying City as an

example. The result was shown in Table 2. The relationship

between the two types of data in the Yellow River Delta with a

difference of more than 200 km2 and the conversion

relationship was shown in Figure 7. In the DLTB-to-LCA

transformation, CLand, WLand, OLand were mainly transferred

out, and FLand and grassland were mainly transferred in. The

FLand increased in the LCA data mainly came from the CLand

and WLand of DLTB. The extra part came from each class,

which mainly came from CLand, WLand and OLand. In the

transformation relationship of more than 200 km2, the

bidirectional relationship were distributed among OLand to

WLand, ULand to WLand, and CLand to GLand. The

differences between the two types of data in the Yellow River

Delta region that is greater than 500km2 were that the CLand in

DLTB was collected as forest land and grassland in LCA; the

WLand was collected as grassland, and OLand was collected as

grassland. With regard to the relationships among the main

agricultural lands in the three major categories of CLand,

GLand and FLand, from DLTB to LCA, the CLand was mainly

transferred out; FLand was mainly transferred in, and the total

amount of GLand remained basically balanced. The main

differences between the two types of data in the coastal four

counties of Dongying were the conversions of OLand, grassland,

and WLand. The transformation relationships between

Shouguang City and Hanting District were mainly the

differences and conversions between OLand/WLand and ULand.

The difference in Zhanhua County was mainly that the CLand

was turned into Gland, and the conversion of CLand to

FLand/grassland was mainly in Wudi County, Qingyun County,

and Bincheng District. The remaining counties were mainly

CLand to FLand. The situation of CLand transferred in and out

in the Yellow River Delta is shown in Figure 8. The difference

and the spatial distribution between the two types of data has

obvious correlation, showing a certain group characteristics.

This may be influenced by a combination of factors such as its

geographical location, natural resources, local cities,

development planning, and economic and land policies.



654

LCA

DLTB 01 02 03 04 10 11 12 20

01 1859.4 22.0 155.0 165.5 13.0 20.6 2.7 62.5

02 18.3 8.89 10.6 1.5 0.34 0.3 0.0 2.6

03 37.4 2.1 76.7 70.6 1.3 18.7 20.1 4.6

04 74.2 0.1 5.3 49.3 0.6 7.6 3.8 5.1

10 10.2 0.4 30.2 37.9 63.8 9.1 2.2 19.1

11 272.7 6.4 89.4 547.6 39.9 1522.1 471.2 219.6

12 283.8 3.1 48.5 489.9 13.5 199.6 108.1 164.9

20 29.29 1.58 43.79 111.62 42.85 26.79 9.22 607.6

Table 2. Source - Direction matrix of LCA and DLTB of

Dongying City (Area Unit: km2)

Figure 7. Classes that the conversion relationships more than

200km2 in the Yellow River Delta

Figure 8. Cultivated Land transferred in and out in the Yellow

River Delta

3.3.3 Application in land supervision: Through in-depth

analysis of the two types of data, the objective and detailed

characteristics of the national geography census data were

highlighted. Combining with the land use hotspots, this paper

proposes that geographical national survey data has several

application points in land supervision. First, the correspondence

between CLand in the two types of data is the most clear. The

emphases is different in the different data system. The LCA data

can be used as a third-party data to monitor and analyse the

transformation of CLand. Through the analysis of the source-

direction of the CLand, we found out the distribution of suitable

land for agriculture and reassessed the suitability of land use. In

particular, administrative department should pay attention to the

phenomenon that a large amount of spontaneous reclamation of

CLand. For example, there were a large number of OLand

(saline land) turning into CLand in the Kenli County Yellow

River estuary. Secondly, the impact of human construction

activities on the land including not only the conversion of non-

construction land to construction land, but also the conversion

of cultivated land from agricultural land to garden/forest land

should also receive enough attention. Of the cultivated land

transferred, 78% were converted into nurseries and 22% of

them were young artificial forests. The CLand converted into

FLand and GLand may be due to the economic construction

activities which could bring more compensation benefits.

Attention should be paid to the construction field, or the

artificial young forest, nursery garden, natural grass around the

town or regularly shaped. Such regions may be the next hot area

of construction. Thirdly, the LCA data can supervise the use of

urban construction land. It can play the role of regulatory in the

progress of construction, construction and planning in line with

the situation and the status approved but no construction. The

large natural grassland within the town, especially the natural

grassland with large area and regular shape, can reflect the use

of construction land within the city, such as the suspected area

that was not used after approval (shown as Figure 4). Through

the comparison of planned land use and actual land cover, it

could be judged whether the construction progress conformed

to the plan. Fourthly, it could track the sources of new

construction land. According to the analysis of the sources of

construction fields, it was found that the construction fields in

the Yellow River Delta mainly coming from ULand (39%),

followed by WLand (21%), then the CLand (16%), shown as in

Figure 9. In the monitoring process, through the analysis of

multiple periods of data transformation, hot spots and key

directions in urban construction could be found.

Figure 9. The sources of construction fields in the Yellow River

Delta



655

4. CONCLUSIONS

The national geography census is application-oriented. Through

the comparison and analysis of the land use data and land cover

classification data, this paper discussed the application of

national geography census data in land supervision. The

objective and meticulous characteristics of the census results

provide a platform for monitoring and supervising the land

census, and its results can also play an important role in land

management. With the comprehensive application of the

national geographical census and the implementation of

normalized geographic national conditions monitoring, he

geographic national condition data will play an increasingly

important role in the management and application of more and

more government departments.

REFERENCES

Bao, G.Y., 2003, Application of Different Methods of

Extracting Varying Information to Dynamic Remote Sensing

Monitoring of Land Use, Bulletin of Surveying and Mapping, 8,

pp. 38-40

Cao, Z.J., Wu, X.Y., Gao, Z.Y., et al, 2013, Land Use Dynamic

Remote Sensing Monitoring Region Partitioning Method Based

on the Development Pressure State: A Case Study of Jinnan

District, Tianjin City, Remote Sensing for Land and Resources,

25(3), pp. 124-129

Chen, J.Y., 2014, Study Notes on Geographic National

Condition Monitoring, Acta Geodaetica et Cartographica Sinica,

41(5), pp. 633- 635

Li, D.R., Sui, H.G., Shan, J., 2012, Discussion on Key

Technologies of Geographic National Conditions Monitoring,

Geomatics and Information Science of Wuhan University, 32(5),

pp. 505-512

Ning, J.S., Wang, Z.T., 2014, Comprehensive report on

development of 2012-2013 surveying and mapping, Science of

Surveying and Mapping, 39(2), pp. 3-10

Qiao, W.F., Sheng, Y.H., Fang, B., et al, 2013, Land Use

Change Information Mining in Highly Urbanized Area Based

on Transfer Matrix: A Case Study of Suzhou, Jiangsu Province,

Geographical Research, 32(8), pp. 1497-1507

Wan, J.B., Zhu, G.R., Peng, Q.G., 2005, Land Use Information

Tupu and Its Application, Geomatics and Information Science

of Wuhan University, 30(4), pp. 355-358

Yang, Y., Xu, L., Yan, P.L., 2014, Urban Land Use/Cover

Classification of Remote Sensing using Random Forests under

the Framework of Conditional Random Fields, Remote Sensing

for Land and Resources, 26(4), pp. 51-55

You, S.C., 2009, Landuse Change Detection Based on CCD

Data of CBERS-02B, Remote Sensing Information, 2, pp. 23-

28

Zhang, F.F., Li, W.Z., Lu, L.Y., 2012, Technologies of

Extracting Land Utilization Infromation Based on SVM Method

with Multi-Window Texture, Journal of Remote Sensing, 16(1),

pp. 73-78

Zhang, X.C., Xiong, X.C.,2008, Research of Spatial Structure

of Land Use Change Based on RS/GIS Technology, Acta

Scientiarum naturalium Universitatis SunYatsen I, 47(3), pp.

117-121

Zhong, K.W., Sun, C.G., Xie, L., 2009, The Dynamic

Monitoring of Land Use Change in Guangzhou Based on RS

and GIS, Journal of Geo-Information Science, 11(1), pp. 111-

116



656

the application research of national ......months each year. the 2014-2015 0.5m resolution aerial...

Documents